Performance Prediction Algorithm Research Articles

Academic Performance Prediction Method of Online Education using Random Forest Algorithm and Artificial Intelligence Methods

In order to improve the teaching quality of online education, the prediction method of students' online academic performance has been studied. First, the learning analysis, artificial intelligence (AI) and other related theoretical concepts are analyzed and introduced. Then, the decision tree of single classification algorithm and the random forest (RF) of ensemble learning algorithm are analyzed, and the academic performance prediction model of online education is constructed by RF algorithm. Finally, the data of education platform is used for empirical analysis to verify the reliability and practicability of the academic performance prediction algorithm of online education. The connotation of learning analysis, the role and elements of learning analysis in the learning process are introduced. The algorithm principle of RF and decision tree is analyzed. By using the idea of information entropy and discretization, the continuous variables are processed to improve the fitting degree of the algorithm. The model is evaluated by empirical analysis, and the test accuracy of several different algorithms is compared. It is found that the prediction accuracy of the RF algorithm is more than 90%, which shows that the prediction method can help teachers and students to carry out better teaching and learning activities, so as to better improve students' ability to master knowledge. It is hoped that the result can provide some reference for the management of students' learning behavior and the optimization of teachers' teaching strategies in online learning activities

Performance Prediction Method for Stream Computing Platform Based on Time Series

As one of the most popular high-performance data processing technology, existing task and resource scheduling strategies for stream computing platforms are suffering from the problem of triggering hysteresis, which seriously affects the cluster performance. To address this problem, the idea of performance prediction based on timeline series data is proposed. Firstly, the performance variation rule of the stream processing platform is analyzed, which provides a basis for proposing the performance prediction model. Secondly, the basic topology paradigm, periodic load prediction model, and real-time throughput prediction model are proposed as the theoretical foundation for performance prediction. Thirdly, the performance prediction algorithm is proposed to predict the variation trend of the processing load and throughput of the cluster. The processing load is predicted periodically while the throughput of the cluster is predicted in a real-time manner. Finally, the performance evaluation algorithm is proposed to evaluate the cluster performance based on the prediction results and trigger the corresponding scheduling strategies in advance. The experimental results showed that the prediction accuracy of the proposed algorithm meets the requirements in practical applications. Meanwhile, the proposed method improves the performance of stream computing performance by triggering the scheduling strategies in advance.

A compounding-model comprising back propagation neural network and genetic algorithm for performance prediction of bio-based lubricant blending with functional additives

Purpose The amount, type and addition conditions of additives of lubricants should be continuously adjusted to obtain appealing performance. To obtain the optimal pretreatment parameters and reduce the cost of time-consuming experiments, the purpose of this paper is to establish an optimal back propagation neural network (BPNN) model combined with genetic algorithm (GA) in this work. Design/methodology/approach Using trimethylolpropane trioleate as the base oil and three types of phosphorus compounds as additives, 25 sets of lubricant formulas were designed regarding lubricant performances of average friction coefficient, average spot diameter, disk wear volume and extreme pressure. The data set was used for training and learning of BPNN and then combined with GA to optimize BPNN with continuously optimization by adjusting various parameters. Findings Comparing prediction data of BPNN with actual test data, correlation coefficients were above 90%, indicating that the model could accurately predict the performance of lubricants. When combined with GA, all performance errors were less than 5%, indicating that BPNN could be optimized by GA to obtain an accurate combined model for prediction of lubricant performance. The best additive formula with excellent performances was obtained from the BPNN–GA model. Originality/value This work developed a new method to study lubricant compounding. The combined model was expected to provide a theoretical basis and guidance for the compounding optimization of lubricant additives with high efficiency and low cost and to expand the scope to practical applications. Peer review The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-05-2020-0165/

Real-time drilling rate prediction method of reaming while drilling based on mechanical specific energy

Reaming while drilling (RWD) is a drilling technique that uses a reamer and a pilot bit to expand the size of an open hole segment to be larger than the inner diameter of an upper casing string while fully drilling, especially in deep wells. This technique reduces the number of trips and is highly efficient and widely used in deep wells, ultra-deep wells, small clearance wells, and side-drilling and complex well conditions. However, in the actual operation process, the mismatching between the pilot bit and reamer often leads to slippage, string vibration, and low-quality borehole.In this paper, based on the mechanical specific energy model and the RWD weight distribution model, combined with the drilling weight, torque, and formation confined compressive strength at the interface of the pilot bit and reamer, a real-time rate of penetration (ROP) prediction model suitable for RWD is derived. Through the analysis of the downhole condition of the RWD operation, reasonable drilling parameters are determined and a set of real-time RWD performance prediction and optimization algorithm is proposed.Studies show that during the RWD operation, the size of the pilot bit and reamer and the shape of the cutter have an enormous influence on ROP. When drilling in different formations, the excellent cooperation among the pilot bit, reamer, and formation is the key to improve ROP and bottom hole assembly stability. How to optimize the bit and reamer to drill with the same ROP and predict the ROP is the focus of this paper.Field logging data test the ROP prediction model and real-time optimization method, and the results indicate that the ROP prediction model has high precision, small error, and high timeliness, which can fully meet the requirements of field engineering. This model can guide the selection of the bit, reamer, and drilling parameters in RWD, which is of considerable significance to RWD.

OP Performance Prediction for Complex Mobile Multiuser Networks Based on Extreme Learning Machine

Due to the complex and variable environments of mobile communication, the mobile multiuser networks become a hot topic. To process active complex event in mobile multiuser networks, it is important to predict the system performance. In this work, the authors consider the multiuser networks which utilizes transmit antenna selection (TAS). We derive novel closed-form expressions for the outage probability (OP) in terms of the Meijer's G-function. Then, a extreme learning machine (ELM)-based OP performance prediction algorithm is proposed. We use the theoretical results to generate training data. We test back-propagation (BP) neural network, locally weighted linear regression (LWLR), wavelet neural network, ELM, and support vector machine (SVM) methods. Compared with wavelet neural network, SVM, BP neural network, and LWLR methods, the Monte-Carlo results shows that the proposed prediction algorithm can consistently achieve higher OP performance prediction results.

BP neural network-based ABEP performance prediction for mobile Internet of Things communication systems

Wireless communications play an important role in the mobile Internet of Things (IoT). For practical mobile communication systems, N-Nakagami fading channels are a better characterization than N-Rayleigh and 2-Rayleigh fading channels. The average bit error probability (ABEP) is an important factor in the performance evaluation of mobile IoT systems. In this paper, cooperative communications is used to enhance the ABEP performance of mobile IoT systems using selection combining. To compute the ABEP, the signal-to-noise ratios (SNRs) of the direct link and end-to-end link are considered. The probability density function (PDF) of these SNRs is derived, and this is used to derive the cumulative distribution function, which is used to derive closed-form ABEP expressions. The theoretical results are confirmed by Monte-Carlo simulation. The impact of fading and other parameters on the ABEP performance is examined. These results can be used to evaluate the performance of complex environments such as mobile IoT and other communication systems. To support active complex event processing in mobile IoT, it is important to predict the ABEP performance. Thus, a back-propagation (BP) neural network-based ABEP performance prediction algorithm is proposed. We use the theoretical results to generate training data. We test the extreme learning machine (ELM), linear regression (LR), support vector machine (SVM), and BP neural network methods. Compared to LR, SVM, and ELM methods, the simulation results verify that our method can consistently achieve higher ABEP performance prediction results.

Predicting Faculty Performance in Higher Education using Machine Learning

Higher education is witnessing significant change in transferring of knowledge and experience from the faculty to the student community on a large scale. Domains requiring a high degree of skill development such as engineering technology education need a place in the university of deeper faculty communication and information percolation process. Despite various measurement and accounting system are prevailing in the current educational system. Known advanced methods for automated evaluation of faculty performance has emanated successfully yet. Here we propose a Machine Learning technique which can interact both with faculty and students to collect feature-oriented parameters of the academic processes prevailing in engineering domain to qualitative and quantitatively access the faculty performance in the modern ways. By using various machine learning methods such as Support Vector Machine, Logistic Regression, Naïve Bayes, Random Forest and Deep learning are used to build the models. The proposed model could identify the faculty’s performance who are likely to improve the course teaching for the students to enhance the teaching – learning process. Model accuracy is used to evaluate the performance of the faculty in the course. To identify the Machine learning algorithm performance F1-Measure and Area under Curve (AUC) value are compared. Evaluation result indicates that Random Forest algorithm is best suitable algorithm for faculty performance prediction for course evaluation through feedback questionaries’ given by the students. It has a 73.1% higher accuracy, 78.7% F1- measure and 0.73 value of AUC when compared to other models. This model is used for decision making support in the higher educational university.

Outage Probability Performance Prediction for Mobile Cooperative Communication Networks Based on Artificial Neural Network

This paper investigates outage probability (OP) performance predictions using transmit antenna selection (TAS) and derives exact closed-form OP expressions for a TAS scheme. It uses Monte-Carlo simulations to evaluate OP performance and verify the analysis. A back-propagation (BP) neural network-based OP performance prediction algorithm is proposed and compared with extreme learning machine (ELM), locally weighted linear regression (LWLR), support vector machine (SVM), and BP neural network methods. The proposed method was found to have higher OP performance prediction results than the other prediction methods.

Application of support vector regression cooperated with modified artificial fish swarm algorithm for wind tunnel performance prediction of automotive radiators

As an important part of automotive thermal management system, finned-tube heat exchangers (FTHEs) are widely used as automotive radiators because of the compact configuration and strong heat transfer capacity. Wind tunnel test is a usual method to investigate the characteristics of heat exchangers, however, its application is limited due to the high cost. Therefore, the authors try to use intelligent algorithm to predict the performance of heat exchangers with different geometrical parameters to decrease the number of wind tunnel tests in this paper. Firstly, the authors choose representative FTHEs with different geometrical parameters, including core width, core height, core thickness, tube depth, fin height and fin pitch, of which the heat transfer capacity are tested on wind tunnel experimental setup. Then the database is established using the obtained experimental data. After data processing, the support vector regression (SVR) model to predict the heat transfer capacity of heat exchangers is established on MATLAB software platform. A modified artificial fish swarm algorithm (MAFSA) of which the effectiveness has been verified by a representative multimodal function is used to improve the accuracy of the SVR model. Thus, the MAFSA-SVR model to predict the heat transfer capacity is established. The root mean square error (RMSE) and the mean absolute percentage error (MAPE) of the MAFSA-SVR model is respectively 1.89 and 1.83%, both of which are better than the SVR model, the Back propagation neural networks (BPNN) model as well as the linear regression (LR) model. It is concluded that there are high accuracy and generalization of the MAFSA-SVR prediction model, by which the heat transfer capacity of a specific heat exchanger can be predicted with a small number wind tunnel tests to reduce exhausting and expensive experimental studies.

Predicting Academic Performance for College Students

Detecting abnormal behaviors of students in time and providing personalized intervention and guidance at the early stage is important in educational management. Academic performance prediction is an important building block to enabling this pre-intervention and guidance. Most of the previous studies are based on questionnaire surveys and self-reports, which suffer from small sample size and social desirability bias. In this article, we collect longitudinal behavioral data from the smart cards of 6,597 students and propose three major types of discriminative behavioral factors, diligence, orderliness, and sleep patterns. Empirical analysis demonstrates these behavioral factors are strongly correlated with academic performance. Furthermore, motivated by the social influence theory, we analyze the correlation between each student’s academic performance with his/her behaviorally similar students’. Statistical tests indicate this correlation is significant. Based on these factors, we further build a multi-task predictive framework based on a learning-to-rank algorithm for academic performance prediction. This framework captures inter-semester correlation, inter-major correlation, and integrates student similarity to predict students’ academic performance. The experiments on a large-scale real-world dataset show the effectiveness of our methods for predicting academic performance and the effectiveness of proposed behavioral factors.

Performance prediction algorithm for autologous PBSC collection in adults and pediatric patients using large volume leukapheresis.

The number of CD34+ cells collected in apheresis procedures depends mainly on the collection efficiency of the device and the blood volume processed. Large volume leukapheresis (LVL) can improve CD34+ cell yield and has previously been investigated using the COBE Spectra device (Terumo BCT, USA). This was a retrospective analysis of LVL performance in patients undergoing continuous mononuclear cell collection (CMNC) using the new Spectra Optia apheresis system (Terumo BCT, USA) at the University Hospital Center, Zagreb, from March 2016 to September 2016. CD34+ cell yield predictability, determined using a customized algorithm, was also assessed. In total, 67 procedures performed in 46 adults and 14 performed in 11 children were included in the analysis. In adults, 30 (65.2%) patients successfully reached their target preapheresis CD34+ cell count on day 1, with a median (interquartile range [IQR]) CD34+ collected cell dose of 4.8 × 106 /kg (2.3-10.6 × 106 /kg). In the pediatric group, 81.8% successfully collected the target CD34+ cell dose on the first day, with a median (IQR) CD34+ collected cell dose of 11.1 × 106 /kg (3.2-16.3 × 106 /kg). The customized algorithm showed a strong and significant linear correlation with actual CD34+ cell dose (P < 0.0001). The results of this study support the use of LVL and the customized prediction algorithm in apheresis procedures. The ability to tailor the procedure to meet the needs of the individual patient may help to minimize the blood volume processed, shorten the duration, reduce the volume of infused anticoagulants, and improve patient comfort.

Academic performance prediction algorithm based on fuzzy data mining

This paper presents an algorithm for prediction of academic performance of students by fuzzy data mining. The fuzzy-trace concept applied to predict the academic performance of the students. An algorithm is proposed in this paper lies with this idea. The fuzzy academic set is generated from the student’s academic data. This is analyzed by the fuzzy-matrix set. The prediction

GWO-BP Neural Network Based OP Performance Prediction for Mobile Multiuser Communication Networks

The complexity and variability of wireless channels makes reliable mobile multiuser communications challenging. As a consequence, research on mobile multiuser communication networks has increased significantly in recent years. The outage probability (OP) is commonly employed to evaluate the performance of these networks. In this paper, exact closed-form OP expressions are derived and an OP prediction algorithm is presented. Monte-Carlo simulation is used to evaluate the OP performance and verify the analysis. Then, a grey wolf optimization back-propagation (GWO-BP) neural network based OP performance prediction algorithm is proposed. Theoretical results are used to generate training data. We also examine the extreme learning machine (ELM), locally weighted linear regression (LWLR), support vector machine (SVM), BP neural network, and wavelet neural network methods. Compared to the wavelet neural network, LWLR, SVM, BP, and ELM methods, the results obtained show that the GWO-BP method provides the best OP performance prediction.

A socio-inspired CALM approach to channel assignment performance prediction and WMN capacity estimation

A significant amount of research literature is dedicated to interference mitigation in Wireless Mesh Networks (WMNs), with a special emphasis on designing channel allocation (CA) schemes which alleviate the impact of interference on WMN performance. But having countless CA schemes at one's disposal makes the task of choosing a suitable CA for a given WMN extremely tedious and time-consuming. There is a conspicuous absence of reliable CA performance prediction metrics to assist in the selection of a high performing CA scheme for a WMN. Popularly used theoretical interference estimation metrics viz., CDALcost, and CXLSwt, have certain flaws which we discuss in this work. We also elucidate the shortcomings of Total Interference Degree (TID) and propose a hypothesis explaining why it is not a reliable CA performance prediction tool. Besides, these metrics are unable to fulfill our ultimate objective of theoretically predicting the expected network capacity of a CA scheme deployed in a WMN, with high confidence. In this work, we propose a new interference estimation and CA performance prediction algorithm called CALM, which is inspired by social theory. We borrow the sociological idea of “a sui generis social reality”, and apply it to WMNs with significant success. To achieve this, we devise a novel Sociological Idea Borrowing Mechanism that facilitates easy operationalization of sociological concepts in other domains. Further, we formulate a Mixed Integer Non-linear Programming (MINLP) optimization model to determine the maximal network capacity of a WMN. Since the MINLP model does not run in polynomial time due to non-linear constraints, we design a heuristic Mixed Integer Programming (MIP) model called NETCAP which makes use of link quality estimates generated by CALM to offer a reliable framework for network capacity prediction. We demonstrate the efficacy of CALM by evaluating its theoretical estimates against experimental data obtained through exhaustive simulations on ns-3 802.11g environment, for a comprehensive CA test-set of forty CA schemes consisting of topology preserving, graph preserving, and graph disrupting CA schemes. We compare CALM with three existing interference estimation metrics and demonstrate that it is consistently more reliable. CALM boasts of an accuracy of over 90% in performance testing, and in stress testing too it achieves an accuracy of 88%, while the accuracy of other metrics drops to under 75%. It reduces errors in CA performance prediction by as much as 75% when compared to other metrics. Finally, we validate the expected network capacity estimates generated by NETCAP, and show that they are quite accurate, deviating by as low as 6.4% on an average when compared to experimentally recorded results in performance testing.

Development of the performance prediction algorithms for cloud services

Main stages of data center service performance prediction were discussed, specifically data monitoring and gathering, calculation and prediction of key indexes and performance index prediction. It was proposed to build data center service performance prediction algorithm based on an analysis of the service transactions index, service resource occupancy index and service performance index. Prediction of the indexes is based on chaotic time series analysis that was used to estimate service transactions index time series trend, the radar chart method to calculate the service resource occupancy index value and weighted average method to calculate service performance index. For performance prediction, it is proposed to use a fuzzy judgment matrix with the service transactions index and service resource occupancy index as input values. It was taken into consideration that service transactions index is usually represented by nonlinear time series and thus the index time series parameters had to be predicted by chaos theory and for the calculation of this index, the estimation procedure of Lyapunov exponent value can be used. The radar chart demonstrates service resource occupancy index estimation of shared storage, mobile storage, memory, computational capability and network bandwidth. The prediction technique was based on the fuzzy nearness category that use input values of transactions index and dynamic changes of the service resource occupancy index.

Support Vector Motion Clustering

We present a closed-loop unsupervised clustering method for motion vectors extracted from highly dynamic video scenes. Motion vectors are assigned to nonconvex homogeneous clusters characterizing direction, size and shape of regions with multiple independent activities. The proposed method is based on support vector clustering. Cluster labels are propagated over time via incremental learning. The proposed method uses a kernel function that maps the input motion vectors into a high-dimensional space to produce nonconvex clusters. We improve the mapping effectiveness by quantifying feature similarities via a blend of position and orientation affinities. We use the Quasiconformal Kernel Transformation to boost the discrimination of outliers. The temporal propagation of the clusters’ identities is achieved via incremental learning based on the concept of feature obsolescence to deal with appearing and disappearing features. Moreover, we design an online clustering performance prediction algorithm used as a feedback that refines the cluster model at each frame in an unsupervised manner. We evaluate the proposed method on synthetic data sets and real-world crowded videos and show that our solution outperforms state-of-the-art approaches.

A health performance prediction method of large-scale stochastic linear hybrid systems with small failure probability

Health performance prediction of a dynamical system aims at determining the probability or possibility that the system state will remain in a permitted area (safe set) or reach a forbidden area (unsafe set) at a future time instance. This paper proposes a health performance prediction algorithm for large-scale Stochastic Linear Hybrid Systems (SLHS) with small failure probability. In the studied SLHS, the continuous variable evolution is described by a set of stochastic linear differential equations, and the discrete state evolution is modeled by a first-order Markov chain. Furthermore, a safe set of the SLHS is described by a permitted area in the hybrid state space. Given an initial condition, a hybrid state evolution algorithm is proposed referring to the execution of stochastic hybrid systems. On this basis, a concept of health degree is introduced to evaluate the health performance of the studied SLHS. Finally, a multicopter with sensor anomalies is studied to validate the availability and effectiveness of the proposed method.

Service demand modeling and performance prediction with single-user tests

Performance load tests of online transaction processing (OLTP) applications are expensive in terms of manpower, time and costs. Alternative performance modeling and prediction tools are required to generate accurate outputs with minimal input sample points. Service Demands (time needed to serve 1 request at queuing stations) are typically needed as inputs by most performance models. However, as service demands vary as a function of workload (load dependent service demands), models that input singular service demands produce erroneous predictions. The alternative, which is to collect service demands at varying workloads, requires time and resource intensive load tests to estimate multiple sample points—which defeats the purpose of performance modeling for industrial use. In this paper, we propose a service demand model as a function of concurrency that can be estimated with a single-user test. Further, we analyze multiple CPU performance metrics (cache hits/misses, branch prediction, context switches and so on) using Principal Component Analysis (PCA) to extract a regression function of service demand with increasing workloads. We use the service demand models as input to performance prediction algorithms such as Mean Value Analysis (MVA), to accurately predict throughput at varying workloads. This service demand prediction model uses CPU hardware counters, which is used in conjunction with a modified version of MVA with single-user service demand inputs. The predicted throughput values are within 9% deviation with measurements procured for a variety of application/hardware configurations. Such a service demand model is a step towards reducing reliance on conventional load testing for performance assurance.

Towards performance prediction of multicore software routers

SummaryRapid advances in network function virtualization technologies have led to the emergence of flexible and scalable embedded data plane processing functions, eg, packet forwarding, on commodity hardware. However, performance prediction of softwarized network functions on such shared resources is challenging but very important for obtaining the full benefits of network function virtualization. This paper addresses the problem of performance prediction for multicore software routers and reveals a key technique to achieve high accuracy. Motivated by observations, we first analyze how many CPU cycles are spent for forwarding a packet in multicore processing systems. Our prediction model of the CPU usage based on cache contention can capture its nonlinear dilation scaled by the number of CPU cores–called dilated CPU consumption (DCC). We validate the accuracy of the DCC model with measured data. On the basis of the DCC model, we develop 2 performance prediction algorithms to predict the maximum throughput of the packet‐forwarding function corresponding to the assigned resources. The first algorithm includes CPU utilization statistics (called DCC‐u), while its simplified version (called sDCC) does not require CPU utilization statistics. We validate the proposed models through the exhaustive set of experiments under various amounts of assigned resources (ie, CPU speeds and the number of CPU cores) and different traffic loads (ie, a large and small number of traffic flows). The results show that DCC‐u and sDCC achieve high accuracy under various amounts of resources and traffic conditions. Remarkably, they improve the precision of estimation from that of the existing techniques by up to 105% to 163%.

Development of a Test Protocol to Measure Uniaxial Fatigue Damage and Healing

Various laboratory testing methods have been developed to characterize the fatigue response of asphalt concrete mixtures. These test methods attempt to simulate in-service conditions or formulate constitutive models. Experiments such as the four-point beam fatigue test have attempted to simulate in-service conditions. The prediction accuracy of such experiments depends on their effectiveness in simulating actual field conditions such as loading, support, stress state, and environment. Constitutive modeling experiments have aimed at measuring fundamental stress–strain relationships so that rigorous constitutive models can be formulated. The results from these experiments are used as input for field performance prediction algorithms. The uniaxial fatigue test is a promising method in this category because of the constant stress state across the specimen section. A few documents have focused on standard test methods for the uniaxial fatigue test; however, there are no AASHTO or ASTM protocols available that include the healing of asphalt concrete mixtures. The main objective of this study was to report on the development of a uniaxial fatigue test protocol that measures fatigue damage and healing of asphalt concrete mixtures. The documented work includes surrogate studies to identify appropriate sample fabrication procedures, gluing materials and procedures, alignment, machine compliance, type of strain wave shape, and strain-control mode of loading. It was found that the use of a gluing jig and 180-mm compaction height was essential to achieve successful mid-specimen failures. In addition, the sinusoidal strain wave shape and on-specimen strain-controlled mode of loading are appropriate test conditions for fatigue damage and healing characterization of asphalt concrete mixtures.